|Publication number||US4910800 A|
|Application number||US 07/240,983|
|Publication date||Mar 20, 1990|
|Filing date||Sep 6, 1988|
|Priority date||Sep 3, 1987|
|Also published as||DE3784930D1, DE3784930T2, EP0305602A1, EP0305602B1|
|Publication number||07240983, 240983, US 4910800 A, US 4910800A, US-A-4910800, US4910800 A, US4910800A|
|Original Assignee||U.S. Philips Corporation|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (6), Referenced by (16), Classifications (9), Legal Events (6)|
|External Links: USPTO, USPTO Assignment, Espacenet|
The present invention relates to a dual branch receiver.
Dual branch receivers well known. One example is described in U.S. Pat. No. 4633315. The interest in dual branch receivers stems primarily from the desire to manufacture as much as possible of a radio receiver as an integrated circuit. A problem in making a receiver circuit by integration is the construction of the filter circuits because it has been long recognised that it is not easy to build a band-pass filter whose pass-band is very narrow. N. F. Barber in an article "Narrow Band-Pass Filter Using Modulation" Wireless Engineer, May 1947 pages 132 to 134 discloses a filter in which an incoming signal is supplied to two branches each including three stages constituted by (1) a modulator for frequency down-converting the incoming signal, (2) a low pass filter for passing the difference component of the modulation and (3) another modulator for frequency up-converting the low pass filtered signal. The modulators in the two branches of the first and third stages are quadrature related with respect to each other. The outputs of the modulators in the third stages are recombined to provide a narrow band-pass filtered signal. Barber discusses the effects of errors resulting from the phase splitting and the differences in gains in the two branches. In the case of using this type of filter in an FM receiver in which after the first stage the modulation is folded about zero frequency, mismatching of gain and deviation from orthogonality between the two signal branches can give rise to an unwanted image being generated. This will result in distortion, and possibly a whistling tone, in the demodulated audio output.
FIG. 1 of U.S. Pat. No. 4633315 shows a dual branch receiver for television signals, comprising first and second branches I and Q, respectively, constituted by first and third mixers and second and fourth mixers. A signal input terminal is coupled to the first and second mixers, each of which also receives a respective one of the in-phase and quadrature phase outputs of an r.f. local oscillator. The r.f. oscillator frequency (fl1) is offset by fo from the input carrier frequency (fc), fo having a value of the order of 100 to 200 Hz. The input signal is mixed down to a baseband, is low pass filtered and in the third and fourth mixers is frequency up-converted using the quadrature related outputs of a second intermediate frequency local oscillator. The in-phase and quadrature-phase signals in the first and second branches are applied to a scanning circuit and a differencing circuit by which the video and sound signals can be obtained.
In U.S. Pat. No. 4633315 gain and phase control are provided to correct for the imbalances between the two paths. Error signals for use in the control are derived from deviations in amplitude and phase of the unwanted image components at the outputs of the summing and differencing circuits. The particular embodiment disclosed in FIG. 1 of the patent uses the picture (or video) carrier signal V as a reference. This carrier signal is applied to a narrowband phase locked loop (PLL) which produces as detected carriers an in-phase carrier and a 90° out-of-phase carrier. These two detection carriers are applied to respective synchronous demodulators which also receive the output V of the difference circuit. The outputs of the synchronous demodulators are low pass filtered to provide d.c. voltages. The d.c. voltage derived using the in-phase carrier from the PLL is applied to an amplitude control circuit which controls the mixing gain of the fourth mixer by amplifying the oscillator mixing signal applied thereto. The d.c. voltage derived using the 90° out-of-phase PLL signal is used to effect phase control by varying the phase quadrature relationship between the two second (I.F.) oscillator mixing signals.
Introducing the small offset frequency fo between the first, r.f., local oscillator frequency fl1 and the input carrier frequency fc enables separation of the unwanted image signal Du to be separated from the desired signal Dw. FIG. 1 of the accompanying drawings shows the signal components at the outputs of the differencing circuit, that is I-Q, assuming that the transmitted carrier fc is unmodulated. In this drawing, the wanted signal component Dw is located at the frequency fl2 +fo, where fl2 is the frequency of the second local oscillator, while the unwanted image component Du is at fl2 -fo. The wanted and unwanted components can be easily distinguished.
The situation at the output of the differencing circuit becomes more complex if the transmitted carrier is frequency modulated. When modulating the signal there are more signal components at the harmonics of the modulating frequency, and their levels are governed by the modulation index employed. A problem arises however if the modulation frequency fm and the offset frequency fo become equal because the harmonics of fm in the wanted signal component Dw will fall on top of the image component Du. As a result it is impossible to detect the image component Du independently of the wanted component Dw in the output of the differencing circuit. If this occurs, the compensation schemes for the deviations in gain and phase between the I and Q branches, for example as described in U.S. patent specification 4633315, will cease to function correctly because they depend on the unwanted image component Du to provide the necessary error signal. Consequently the demodulated output from the receiver will suffer from distortions.
An object of the present invention is to avoid distortions in the receiver output.
According to the present invention, in a dual branch receiver comprising a signal input terminal coupled to first and second branches, the first branch including first and third mixers, the second branch including second and fourth mixers; a first local oscillator for producing a first local oscillator signal which is supplied to the first and second mixers for frequency down-converting an incoming signal applied to the input terminal, the first local oscillator frequency being offset from the carrier frequency of the incoming signal; a first 90° phase shifter for providing a quadrature phase relationship between one of the signals applied to one of the first and second mixers and the corresponding signal applied to the other of the second and first mixers; a second local oscillator for producing a second local oscillator signal which is applied in quadrature to the third and fourth mixers for frequency up-converting the quadrature related signals produced by the first and second mixers; and a summing circuit and a differencing circuit each having inputs coupled to outputs of the third output distortion is reduced by including fourth mixers, and means for frequency modulating the first local oscillator signal with a wobbling frequency which is less than the offset frequency.
The present invention is based on recognition of the fact that by wobbling the first local oscillator frequency, apart from two instants in time per period of the wobbling signal when the amplitude of the wobbling signal becomes zero, the error signal components Du are separated from Dw by a frequency margin of (2|fo-fw|), and can thus be detected independently from the output of the differencing circuit.
The wobbling frequency fw is low and may be of the order of say 100 Hz. Such a small variation of the offset frequency has an advantage over a large variation of the offset frequency, because there is no need to provide a significant extension in the bandwidth of low pass selectivity filters in the output circuits of the first and second mixers. The wobbling frequency may be produced by a sinusoidal signal, a triangular signal or a ramping signal.
In an embodiment of the present invention the first local oscillator is a voltage controlled oscillator having a frequency control input. The frequency modulating means includes means for applying a wobbling signal voltage to a d.c. frequency control voltage applied to the frequency control input of the voltage controlled oscillator.
If desired the receiver may include automatic gain control means and automatic phase control means for compensating for gain and phase differences resulting from transmission differences in the first and second branches.
The present invention will be explained and described, by way of example, with reference to the accompanying drawing.
FIG. 1 illustrates the signal components in the output of a differencing circuit of a receiver not having wobbling of the first local oscillator frequency and the incoming carrier being unmodulated,
FIGS. 2 and 3 illustrate respectively signal components at the output of the differencing circuit in the presence of modulation with a single sinusoidal frequency fm, where fm<fo (FIG. 2) and fm=fo (FIG. 3),
FIG. 4 is a block schematic circuit diagram of a receiver in accordance with the present invention,
FIG. 5 illustrates the relationship in the frequency plane between the desirable and image signal components, Dw and Du, respectively, when the offset frequency is varying and the modulation (fm) substantially equals the nominal offset frequency, fo,
FIGS. 6 and 7 are comparative curves of distortion versus the modulation index, β, for audio frequencies of 300 and 800 Hz and without and with frequency wobbling, and
FIG. 8 is a block schematic diagram of a substantially complete dual branch receiver made in accordance with the present invention.
In the drawings corresponding reference numerals have been used to indicate similar features.
Referring initially to FIG. 4 the dual branch receiver comprises first and second branches 12, 14 which are connected to an input terminal 10. The first branch 12 comprises a first mixer 16 in which the input signal is frequency down-converted, a low pass filter 18 which selects the down-converted component in the output from the first mixer 16 and a third mixer 20 in which the output signal from the filter 18 is remodulated or frequency up-converted. The second branch 14 comprises a second mixer 17 which produces a quadrature related frequency down-converted signal, a low pass filter 19 which selects the down-converted component in the output from the second mixer 17 and a fourth mixer 21 which produces a quadrature related frequency up-converted signal. The signal branch 12 is termed the in-phase branch I and the signals at the outputs of circuit elements 16, 18 and 20 are referenced I1, I2 and I3, respectively. The signal branch 14 is termed the quadrature phase branch Q and the signals at the outputs of circuit elements 17, 19 and 21 are referenced Q1, Q2 and Q3, respectively. The signals I3 and Q3 are applied to summing and differencing circuits 22, 24, respectively, which provide SUM and DIFF signals. A frequency discriminator 26 is connected to the output of the circuit 24 to produce for example an audio output.
The first and second mixers 16, 17 are provided with quadrature related first local oscillator signals fl1 which are offset slightly, say by 100 Hz, relative to the carrier frequency fc of the input signal by a first local oscillator 28. It is of course possible to shift the input signal in the branch 14 by 90° and supply local oscillator signals of the same phase to both mixers 16, 17. A second local oscillator 30 produces quadrature related second local oscillator signals fl2 at an intermediate frequency of say 100 kHz which are applied to the third and fourth mixers 20, 21.
A source 32 of wobbling signals is connected to the local oscillator 28 to frequency modulate the local oscillator signals. The wobbling frequency in this example is nominally 20 Hz and has a maximum deviation of 100 Hz. The deviation is controlled by the peak amplitude Aw of the wobbling signal. A low wobbling frequency is selected so that the bandwidth of the low pass filters 18, 19 remains substantially unchanged or is not significantly extended. The wobbling signal can be a sinusoidal signal, a triangular signal or a ramping signal.
In order to understand the reason for wobbling the first local oscillator frequency, initially situations will be described in which there is no wobbling of the frequency fl1.
The situation which happens when there is no modulation of the carrier frequency has been described already with reference to FIG. 1 and accordingly in the interests of brevity the description will not be repeated.
FIG. 2 extends the discussion of FIG. 1 by considering the situation at the output of the differencing circuit 24 when the transmitted carrier is frequency modulated by a sinusoidal signal of frequency fm, fm being less than the offset fo, such that the input signal rt is
rt=A sin (2πfct+β sin (2πfmt))
where A is the peak amplitude of the received signal and β is the modulation index.
The resultant signal components at the output of the differencing circuit 24, in the presence of deviations in gain and phase in the branches 12, 14, can be represented in the frequency domain shown in FIG. 2. Because of the modulation, there are more signal components at the harmonics of the modulating frequency fm, and their levels, indicated by the relative heights of the arrows in FIG. 2, are governed by the modulation index β employed. For ease of illustration, only three harmonic components are shown in FIG. 2 for the desired signal Dw and the unwanted image Du. It is evident from an examination of FIG. 2 that as long as the modulating frequency fm differs from the offset frequency fo, the Dw and Du signals will be separated by a frequency margin of (2|fo-fm|) and in consequence can be separately distinguished.
FIG. 3 illustrates the situation when fm=fo. Harmonics of fm in Dw will coincide in the frequency domain with the image Du. As a result, it is impossible to detect the image Du independently of the signal Dw from the DIFF output shown in FIG. 3. If this occurs, the compensation schemes for the deviations in gain and phase between the I and Q channels (or the branches 12, 14) as described in U.S. Pat. No. 4633315, or later herein with respect to FIG. 8 of the accompanying drawing will cease to function correctly because they depend on the image Du to provide the necessary error signal. Consequently, the demodulated audio output will suffer from severe distortions, which are unacceptable.
In the receiver made in accordance with the present invention, the image Du is made separable from Dw for substantially all modulating frequencies fm by continuously varying the offset frequency fo. This is achieved by wobbling the first local oscillator frequency fl1 as described already with reference to FIG. 4.
FIG. 5 illustrates the relationship in the frequency domain between Dw and Du for the case shown in FIG. 3 but with wobbling of the offset frequency fo.
From FIG. 5 it can be observed that because fl1 is changing its frequency at a rate governed by fw, then the Du components are only completely coincident with the Dw components at two instants in time per period of the wobbling signal, i.e. when the amplitude Aw of the wobbling signal becomes zero. At other instants in time, the error signal components Du are separated from Dw by a frequency margin of (2|fo-fw|), and can thus be detected independently from the DIFF output. By a proper choice of the wobbling signal, and the time constants of the correcting loops described with reference to FIG. 8, it is possible to minimize the number of instants that Du is "drowned" by Dw which is normally much larger in amplitude.
In one embodiment fl1 is frequency modulated by a wobbling signal of frequency fw of 20 Hz, with a maximum frequency deviation of 100 Hz. The nominal offset frequencies fo adopted are 300 Hz and 800 Hz, and the corresponding modulating frequencies fm are 300 Hz, and 800 Hz, respectively. In both cases, the frequencies of the modulating sinusoidal signal are made equal to the nominal offset frequencies. The measured distortions in percent expressed as a function of the modulation indices β are shown in FIGS. 6 and 7. In the experiments when the offset frequency fo remains fixed, severe distortions occur in the demodulated audio output curves (1). However, when the local oscillator frequency is wobbled at the output frequency fo distortions have greatly been reduced, curves (2).
The wobbling signal used need not be restricted to a sinusoidal waveform but it could be any convenient signal which would not cause a significant extension in bandwidth of the two low pass selectivity filters 18, 19 (FIG. 4). Also the additional wobbling signal will give rise to a demodulated output, and this has to be filtered out at the output of the frequency discriminator 26 (FIG. 4). In the above example, fw is chosen to be 20 Hz, so that it can be filtered out by the bandpass filter normally employed after the discriminator. This filter is needed for filtering the normal audio signal, which for mobile radio has a lower cutoff frequency of 300 Hz. In this case, no additional filtering is needed to remove the demodulated signal term caused by the wobbling signal.
Although only the DIFF output has been referred to in the above description, the same happens with the SUM output of FIG. 4.
FIG. 8 illustrates a complete receiver made in accordance with the present invention, which receiver has means for deriving control signals for correcting gain and phase deviations.
The first and second branches 12, 14 are as described with reference to FIG. 4. The first local oscillator 28 is shown connected to a 90° phase shifter 29 having quadrature outputs. The second local oscillator 30 is shown connected to an adjustable phase shifter 31 which is responsive to a control signal to vary the relative phase of its outputs to compensate for phase deviations. An adjustable gain amplifier 34 is connected in the first branch 12 at the output of the third mixer 20 and a fixed gain amplifier 35 is connected in the second branch 14 at the output of the fourth mixer 21. The adjustable gain amplifier 34 is responsive to a control signal to equalise the gain in the branch 12 with that of the branch 14.
The image signal in either of the outputs of the summing circuit 22 (SUM) or the differencing circuit 24 (DIFF) could be used as a measure of the amplitude of deviations in gain and phase. For illustration the DIFF signal has been chosen for his purpose. The DIFF signal is bandpass filtered in a filter 60, has its amplitude adjusted, if required, in a variable gain amplifier 62 and is self-multiplied with itself in a first multiplier 46. D.C. signals produced by the self-multiplication are blocked by a high pass filter 47 while signals centerd at 2.fl2 are filtered away by a low pass filter 48. In practice the filters 47, 48 are realised as a capacitance-resistance (CR) high pass network and as a resistance-capacitance (RC) low pass network, respectively. The filtered product of the self-multiplied DIFF signal includes a signal centred at 2fo which is used for deriving the necessary control signals.
A second, contemporaneous operation comprises multiplying the DIFF signal by the SUM signal in a second multiplier 49. For this purpose the bandpass filtered DIFF signal is shaped in a limiter 63 and applied via the in-phase output of a broadband phase locked loop 54 to one input of the second muliplier 49. The SUM signal is bandpass filtered in a filter 61, shaped in a limiter 64 and is applied to a second input of the second multiplier 49. D.C. signals produced by the multiplication in the multiplier 49 are blocked by a second high pass filter 50 while signals centred at 2.fl2 are filtered away by a second low pass filter 51 to leave signals centred at 2fo.
The signals centred at 2fo in the outputs of the filters 48, 51 are multiplied together in another multiplier 52 and the product is low pass filtered in a filter 53 to provide a d.c. signal which is applied to the amplifier 34 as a gain control signal.
Phase control is applied by shifting the phase of either the DIFF or SUM signal by 90° (in the illustrated example the DIFF signal is phase shifted) and is multiplied by the other signal. The 90° output of the phase locked loop 54 provides the 90° phase shifted DIFF signal at one input of a third multiplier 55, the other input of which is connected to receive the waveform shaped SUM signal. The product is filtered in a third high pass filter 56 and then in a third low pass filter 57 again to provide signals centred at 2fo. The outputs of the low pass filters 48, 57 are multiplied in a further multiplier 58 and the output is low pass filtered in a filter 59 to provide a d.c. phase control signal which is applied to the phase shifter 31.
The source 32 of wobbling signals is arranged to act on the first local oscillator 28 which in the illustrated embodiment is implemented as a voltage controlled oscillator connected to a 90° phase shifter 29. An automatic frequency control of the first local oscillator is necessary to maintain a certain accuracy in the specified offset frequency fo. In practice, the value of fo is normally governed by a d.c. reference signal. In order to wobble fo, it is only necessary to introduce the wobbling signal in addition to this d.c. reference signal. The combined signal is then used to control automatically the frequency of the first local oscillator 28.
In order to implement this automatic frequency control, the output from an offset frequency detector, such as the output of a frequency discriminator in the PLL 54, is applied to one input 68 of a comparator 70. A d.c. reference signal to provide for the nominal offset frequency is applied to one input 72 of a summing circuit 74. A source of a wobbling signal (not shown) is applied to a second input 76 of the summing circuit 74. The summing circuit 74 has an output connected to a second input 78 of the comparator 70. The output from the comparator 70 comprises an afc signal for the first local oscillator 28.
In the event of the signal of the first local oscillator being synthesised from a frequency synthesised source then a nominal offset frequency can be built in. Additionally the wobbling signal can be introduced at a frequency or phase modulation part of the synthesizer, the output of which forms an input of a voltage controlled oscillator of the frequency synthesizer.
The source of wobbling signals may comprise a simple oscillator since its output signal can be of a low quality in terms of signal purity and accuracy in frequency. Alternatively it could be derived from the second local oscillator 30 by dividing its output frequency down using a frequency divider followed by an appropriate wave shaping network.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US3624507 *||Apr 21, 1967||Nov 30, 1971||Masao Fukata||Communication system of a cue signal or signals|
|US4313219 *||Mar 11, 1980||Jan 26, 1982||Siemens Aktiengesellschaft||Receiver for high frequency electromagnetic oscillations having a frequency readjustment|
|US4580101 *||Apr 4, 1984||Apr 1, 1986||Multitone Electronics Plc||FM demodulators with local oscillator frequency control circuits|
|US4633315 *||Mar 23, 1984||Dec 30, 1986||U.S. Philips Corporation||Receiver for RF-signals comprising a pair of parallel signal paths|
|US4682117 *||Apr 30, 1986||Jul 21, 1987||U.S. Philips Corporation||Quadrature demodulation data receiver with phase error correction|
|US4736390 *||Oct 15, 1986||Apr 5, 1988||Itt Avionics, A Division Of Itt Corporation||Zero IF radio receiver apparatus|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US5146338 *||Aug 6, 1990||Sep 8, 1992||Thomson Consumer Electronics, Inc.||Fixed rf agc of a television tuner for fm reception in a television receiver|
|US5212825 *||Nov 9, 1990||May 18, 1993||Litton Systems, Inc.||Synthetic heterodyne demodulator circuit|
|US5418489 *||Oct 8, 1993||May 23, 1995||Motorola, Inc.||Method of recovering a frequency modulated signal|
|US5428836 *||Oct 14, 1993||Jun 27, 1995||Motorola Inc.||Radio receiver for forming a baseband signal of time-varying frequencies|
|US5710997 *||Jan 24, 1996||Jan 20, 1998||Sony Corporation||Superheterodyne receiver apparatus|
|US5898906 *||Oct 29, 1996||Apr 27, 1999||Gec Plessey Semiconductors, Inc.||System and method for implementing a cellular radio transmitter device|
|US5901349 *||Dec 11, 1996||May 4, 1999||Matra Communication||Mixer device with image frequency rejection|
|US6148039 *||Jan 30, 1998||Nov 14, 2000||Scientific-Atlanta, Inc.||System and method for harmonic interference avoidance in carrier recovery for digital demodulation|
|US6993295 *||Jan 8, 2002||Jan 31, 2006||Intel Corporation||Weaver image reject mixer with fine resolution frequency step size|
|US7362825 *||Apr 6, 2005||Apr 22, 2008||Sony Corporation||Receiving device and integrated circuit for reception|
|US7567790||Apr 6, 2005||Jul 28, 2009||Sony Corporation||Receiving device and integrated circuit for reception|
|US20030129959 *||Jan 8, 2002||Jul 10, 2003||Gilmore Robert P.||Weaver image reject mixer with fine resolution frequency step size|
|US20050190862 *||Apr 6, 2005||Sep 1, 2005||Sony Corporation||Receiving device and integrated circuit for reception|
|US20050208937 *||Apr 6, 2005||Sep 22, 2005||Sony Corporation||Receiving device and integrated circuit for reception|
|WO1993014578A1 *||Nov 6, 1992||Jul 22, 1993||Motorola Inc.||Radio receiver for forming a baseband signal of time-varying frequencies|
|WO1995010879A1 *||Aug 22, 1994||Apr 20, 1995||Motorola, Inc.||Method of recovering a frequency modulated signal|
|U.S. Classification||455/316, 455/304, 455/315, 455/303, 455/209|
|International Classification||H03D7/16, H04B1/26|
|Mar 29, 1989||AS||Assignment|
Owner name: U.S. PHILIPS CORPORATION, NEW YORK
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST.;ASSIGNOR:CHUNG, KAH-SENG;REEL/FRAME:005044/0090
Effective date: 19890302
|Aug 26, 1993||FPAY||Fee payment|
Year of fee payment: 4
|Sep 2, 1997||FPAY||Fee payment|
Year of fee payment: 8
|Oct 9, 2001||REMI||Maintenance fee reminder mailed|
|Mar 20, 2002||LAPS||Lapse for failure to pay maintenance fees|
|May 14, 2002||FP||Expired due to failure to pay maintenance fee|
Effective date: 20020320